happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

missing __func__, move imatrix flag set

This commit is contained in:
ddh0 2026-02-11 20:57:16 -06:00
parent 2769f35207
commit ea8da0503c
1 changed files with 6 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
src/llama-quant.cpp
View File

@ -937,13 +937,13 @@ static void llama_model_quantize_impl(const std::string & fname_inp, const std::
tensor_size/1024.0/1024.0, tensor_size/1024.0/1024.0,
new_size/1024.0/1024.0, new_size/1024.0/1024.0,
ggml_type_name(new_type)); ggml_type_name(new_type));
if (!will_require_imatrix && tensor_requires_imatrix(params, tensor, new_type)) {
will_require_imatrix = true;
}
} else { } else {
new_size = tensor_size; new_size = tensor_size;
LLAMA_LOG_INFO("size = %8.3f MiB\n", new_size/1024.0/1024.0); LLAMA_LOG_INFO("size = %8.3f MiB\n", new_size/1024.0/1024.0);
} }
if (!will_require_imatrix && tensor_requires_imatrix(params, tensor, new_type)) {
will_require_imatrix = true;
}
total_size_org += tensor_size; total_size_org += tensor_size;
total_size_new += new_size; total_size_new += new_size;
continue; continue;
@ -1072,7 +1072,9 @@ static void llama_model_quantize_impl(const std::string & fname_inp, const std::
LLAMA_LOG_INFO("%s: model size = %8.2f MiB (%.2f BPW)\n", __func__, total_size_org/1024.0/1024.0, total_size_org*8.0/ml.n_elements); LLAMA_LOG_INFO("%s: model size = %8.2f MiB (%.2f BPW)\n", __func__, total_size_org/1024.0/1024.0, total_size_org*8.0/ml.n_elements);
LLAMA_LOG_INFO("%s: quant size = %8.2f MiB (%.2f BPW)\n", __func__, total_size_new/1024.0/1024.0, total_size_new*8.0/ml.n_elements); LLAMA_LOG_INFO("%s: quant size = %8.2f MiB (%.2f BPW)\n", __func__, total_size_new/1024.0/1024.0, total_size_new*8.0/ml.n_elements);
if (!params->imatrix && params->dry_run && will_require_imatrix) { if (!params->imatrix && params->dry_run && will_require_imatrix) {
LLAMA_LOG_WARN("%s: WARNING: dry run completed successfully, but actually completing this quantization will require an imatrix!\n"); LLAMA_LOG_WARN("%s: WARNING: dry run completed successfully, but actually completing this quantization will require an imatrix!\n",
__func__
);
} }
if (qs.n_fallback > 0) { if (qs.n_fallback > 0) {